Dell Tech World 2025 Delivers Real Enterprise AI — At Full Speed

Forbes

2 days ago

Dell Technologies COO Jeff Clarke presents during the second day of Dell Technologies World 2025.
AI-driven change is coming faster than ever, with no signs of slowing down. That's what I encounter every day as I talk with business leaders across the tech landscape, and it was a message that came through loud and clear at Dell Technologies World in Las Vegas last month. Based on what we heard from the executives on stage and what Michael Dell said when my business partner and I interviewed him, Dell Technologies is embracing this trend — harnessing AI to transform its own business at the same time it helps customer companies implement AI across many different use cases.
A year ago, the company used Dell Tech World to launch its AI Factory offering in partnership with Nvidia. Now it has more than 3,000 customers of all sizes for AI Factory, which is a major driver for what Michael Dell calls 'a big rearchitecture going on with AI.' From my perspective, Dell Tech World 2025 was a key progress report from one of the companies that's central to a broader redefinition of edge computing, AI infrastructure and — maybe most critically — the process of making enterprise AI real.
(Note: Dell Technologies is an advisory client of my firm, Moor Insights & Strategy.)
Fortunately, Dell executives did a lot at DTW 2025 to show that the big splash around AI Factory in the past year can't be dismissed as marketing hype. The company has helped those 3,000 customers launch myriad AI projects; according to the company's stats, about three-quarters of those initiatives have met or exceeded expectations, with average improvements of 20% to 40% in key metrics. On top of that, AI Factory — available in your choice of AMD or Intel flavors — is touted as 60% more cost-effective than using a public cloud for AI.
That comparison to cloud performance tied in with an important prediction from Michael Dell. He thinks that 75% of all data will be created and processed at the edge sometime in the near future, and that, crucially, 'AI will follow the data, not the other way around.' In this context, he's focused on decentralization, low latency and hyperefficiency for enterprise AI. While I think that the big cloud service providers would argue on behalf of their own efficiency, they are understandably singing a different tune when it comes to decentralization.
As a recovering product person, I love diving into the gnarly details of processor specs, server configurations and so on, and the presentations at DTW 2025 — let alone the drill-down breakout sessions — did not disappoint. From the stage, Michael Dell cited the example of a datacenter the company is deploying that uses 110,000 GPUs and generates trillions of tokens. Indeed, the exponential growth of tokens, compute, data and so on was a recurring theme.
While the scale of the company's efforts is impressive enough, it doesn't entirely surprise me coming from Dell, which is one the top server vendors in the world, and by far the biggest in storage. What does impress me is that Michael Dell and his lieutenants seem more dedicated than ever to giving customers an 'easy-button' way to implement all of this scale and complexity. (Hold that thought, because I'll return to it later.)
Apparently, this approach is working in the real world for customers. For example, there was a presentation from Larry Feinsmith, who heads global technology innovation for JPMorgan Chase. He touted the bank's 30-year partnership with Dell and talked about the importance of Dell technology for helping the bank move $10 trillion daily. The figures were again impressive, though this is the sort of thing you come to expect when a giant customer of a giant tech company takes the stage at an event like this. It got more interesting when Feinsmith got specific about AI. His firm has rolled out an in-house LLM to more than 200,000 people internally — the largest rollout of that kind ever, as far as the bank and Dell are aware. Feinsmith also quoted one of his colleagues as saying that generative AI is 'as transformative as any technology we've seen in the past few hundred years.'
While I agree about the transformative potential, it's even more important that Dell understands the open secret about AI. 'It's not just about the technology,' the CEO said. 'It's about how enterprises can create and capture value.' In that vein, he expects a $1 trillion investment globally in AI in the next five years that will add something like $15 trillion to global GDP. Not just incidentally, he thinks that the number of Dell AI Factories will increase from thousands now to millions by then.
As expected, the company also announced a slew of new hardware to populate AI Factories and other datacenters. This includes both air- and liquid-cooled versions of PowerEdge XE9780 and XE9785 servers, which support up to 192 Blackwell Ultra GPUs from Nvidia, with customization options up to 256 of those GPUs. Advances in cooling — including direct-to-chip liquid cooling — were promoted as saving as much as 60% on cooling costs. (Considering both the direct costs of datacenter power and the longer-term potential for electricity scarcity for power-hungry AI compute, it's no accident that Lenovo, HPE and Dell all talk regularly about the state of their cooling technologies.) There are also PowerEdge XE9712 and (soon) XE7745 servers tooled specifically for AI functions including training, inference and agentic use cases. These are complemented by new high-density switches from Dell and Nvidia that handle either Ethernet or InfiniBand networking.
All of this hardware is backed up by an extensive suite of services. This includes consulting for every phase of AI projects, from strategy through implementation to the fine points of operating and scaling production deployments — not to mention related services for security and resilience. These go along with training, certification and support at every relevant stage, as well as a list of managed services specifically to support AI Factory and the Dell Pro AI Studio. In short, Dell has invested in supporting customers across the full lifecycle of this infrastructure to help them implement AI more simply and turn ideas for AI use cases into reality more quickly. And the real music to my ears was the announcement of the new Data Management Services. Getting the data right has always been one of those eat-your-vegetables chores that you must do for technology to work well. Absolutely everything I've seen over the past two years of the generative AI revolution confirms that getting the data right is more important than ever, so I believe these services will be real difference-makers for many of Dell's clients.
Rather than go into more detail about the announcements for the Dell AI Data Platform, I'll point you to my colleague Matt Kimball's article from last week. The short version is that Dell is going to great lengths to supply — either directly or with partners — the servers, storage, networking, data management and software for large-scale enterprise AI deployments, with services to back it up.
One of the key insights from Kimball's post is that Dell is both building upon and transcending its history as an infrastructure vendor to provide business customers of all sizes with complete AI solutions. Now, if you've spent time around tech marketing like I have, you're probably as allergic as I am to empty claims about 'end-to-end solutions' — an often lazy term that's been badly overused since at least the dot-com bubble. (Note that Dell itself is savvy enough to avoid this.) But sometimes a company really does take an approach that actually solves the problems of its business customers, and does it in a comprehensive way. That's the position that Dell has put itself in now for enterprise AI.
Just think about the company's strengths. Decades of building and selling big-iron hardware? Check. A world-class supply chain to deliver equipment and services? One of the best. Offerings beyond the datacenter to the network edge? Yes. Fleets of client PCs? Dude, it's Dell.
All of the advances that Michael Dell, his executives and their customers talked about at the show require replacing and expanding aging infrastructure with AI-enabled Dell technology. When Nvidia CEO Jensen Huang joined the event by video, he talked about the entire tech stack being reimagined as a result of 'unquestionably the single biggest platform shift' in the history of IT. Put another way, AI inherently drives a high level of IT transformation and consolidation because it requires new architecture. And Dell is standing by with flexible options to meet that need and to run 'any workload, anywhere.'
Capping it all off — and getting back to Kimball's point about solutions — are the data management and other service offerings mentioned earlier that make this extremely digestible for customers across many different verticals who are facing many difficult questions about AI. And the company's affinity for effective partnerships means that it can move faster and cover even more use cases. These offerings are exactly what is needed to scale AI rapidly, within one company and across a wide customer base.
To me, the current moment in AI is like the second generation of the Internet, when things were growing and changing so quickly — in technical terms and in business terms — that empires could rise and fall in no time flat. If anything, the rate of change today is even faster than that. Michael Dell thinks so, and I agree.
When my business partner Daniel Newman and I interviewed him onstage at the event, Dell pointed out the trend that 'businesses have been succeeding or failing at a faster rate because of increasing change.' He thinks that the AI era 'will separate the winning and losing companies faster than ever before.' He also noted that one of the challenges for incumbents across industries is the rise of new 'born-AI companies' that have always incorporated AI and therefore don't have the same baggage as the incumbents that need to restructure everything.
That said, a vendor doesn't get to be the size of Dell Technologies by catering only to early adopters. Now it's out there in the field, helping large, medium and small companies with various levels of data-readiness and AI-readiness take the necessary steps to embrace enterprise AI. To provide an example from outside of tech or the early-adopting financial services industry, the event included a presentation from Seemantini Godbole, the CIO of Lowe's. She talked about the big retailer's digital transformation, starting with a website crash on Black Friday 2018 and extending to its current embrace of cutting-edge AI. Each of 1,700-plus stores has a sort of mini-datacenter to collect and process data at the edge (supporting Michael Dell's earlier point), and each store employee gets to use 'basically a ChatGPT for home improvement' to help customers find what they need. And there are many more use cases already addressed or in the works, from the retail floor all the way to corporate HQ. Lowe's is not a company that's going to build its own LLMs, but it has seen the wisdom of getting ahead of the curve with AI in the context of serving its customers better.
When we talked, I asked Michael Dell to reflect on how his company — after 40-plus years in business and with so much success under its belt — has embraced AI to transform itself. His focus on the human side of the equation stood out to me. 'This is ultimately a change management activity,' he said. You have to explain to your people what's going on, why it's important. You figure out 'What does it actually mean to each person and each group and each activity inside the company?' And then you apply all of that to figuring out how AI helps the company better carry out its mission and be more successful.
Let's grant that this is a critical moment in the history of technology. To capitalize on it, Dell has to keep making enterprise AI real for thousands — tens of thousands — of other businesses, removing barriers around cost and data management and complexity in general so that AI becomes truly applicable for all of those organizations. I believe that the company can likely do it, but it's unquestionably a tall order.
I now think that the true competitors to Dell are the hyperscalers, so Dell needs to think about how to add end-to-end AI platforms, or at least string together the partners to do it, on-prem. In other words, I believe the industry now needs an on-prem version of Vertex AI, Bedrock or Sagemaker — and Dell is well-situated to do it.
Michael Dell himself believes that 'We're a long way from the whole world being reimagined [with AI], and that's going to take years to occur.' In fact, he thinks we're in the early stages of AI's impact. 'I really think we're still in the first 10% of this curve adoption,' he said. 'Because you do see some companies that are going ahead, but . . . a lot of customers, they're not exactly sure what to do with all this, how to do it, when, where to start.' His company is helping them work through all the knotty issues of data handling, storage, compute, privacy, security, model selection and so on to address the relevant use cases, from 'the bite-sized pieces of it that work at the level of a retail store or inside a telco cellular base station or inside a small restaurant or a gas station in your PC, all the way up to these massive training models.' There's no end to the details.
Mind you, other very smart, very disciplined, very well-run companies are also applying themselves to some overlapping set or subset of these activities. As I said above, out of all of them I suspect that Dell Technologies may be on a collision course with the hyperscalers, which have a vested interest in doing enterprise AI their way. Dell is unique in that it spans all the way from the client computing enterprise edge to the CSP to the enterprise data center. The hyperscalers don't do that — but AWS, Microsoft, Google and the other companies in that sector have strengths of their own.
Whatever the case, the company's founder seems primed to once again 'Play nice but win,' as the title of his bestselling book advises. He calls the current moment 'maybe the most exciting time ever in our industry that I've experienced.' He sees a 'window of opportunity to jump ahead' and doesn't want to miss it, for his company or his customers.
I think it's still too early to declare many long-term winners and losers in the AI race. Nvidia's an easy call, but there are too many unknowns to say exactly how Dell and AWS and IBM and OpenAI and Microsoft and Anthropic and Meta and the rest — so many different companies coming at these challenges from different angles — will stack up. But based on everything I've seen, including all the strong evidence coming out of Dell Tech World this year, I have to think that Dell is a smart bet.